building AI
Repositories
peytontolbert/F5TTS-FASTAPI
A FastAPI service for text-to-speech synthesis using the F5-TTS model. Includes authentication token
peytontolbert/SOPLibrary
Library of Software Engineering Standard Operating Procedures (SOP)
peytontolbert/simple_transformer
A simple transformer model in PyTorch - Day 11 of 365 Days of Repos
peytontolbert/repository-skill-miner
This is a skill builder that ingests repositories for contract-driven skill artifacts, enhancing LLM agents with verifiable and traceable code.
peytontolbert/Griffin
Griffin: Mixing Gated Linear Recurrences with Local Attention for Efficient Language Models
peytontolbert/model_bridge
Run certain legacy models from one up-to-date runtime instead of manually switching through environments.
peytontolbert/DynamicAgent
Dynamic Agent, empowered by a Graph RAG LLM
peytontolbert/f5-tts
A repository for voice cloning using the F5-TTS model
peytontolbert/model-stack
A modular model stack that cleanly separates the concerns required to train, serve, and evaluate large language models.
peytontolbert/Research_Library
knowledge system that indexes code repositories and research papers. 1M papers available on HF
peytontolbert/continuous-learning-assistant
The Continuous Learning Assistant is an intelligent chat-based system designed to interact with users, process their inputs, retrieve relevant knowledge, generate meaningful responses, collect feedback, and continuously improve its performance through fine-tuning.
peytontolbert/AI_Atlas
A standalone, interactive dependency map of artificial intelligence organized as a family-level ontology rather than a list of products or model releases.
peytontolbert/agent_kernel_lite
Public code for BitNet model inference in wasm (runs on iphone)
peytontolbert/peytontolbert
peytontolbert/distill-kit
A programming‑focused PyTorch toolkit for building and distilling decoder models. Includes Mixture‑of‑Experts (MoE), pluggable memory backends, composition utilities, and checkpoint compatibility.
peytontolbert/ConfidenceTransformer
This repository contains the implementation of a Confidence Transformer model based on GPT-2. The model is designed to provide confidence scores for its predictions, detect out-of-distribution (OOD) inputs, and perform language modeling tasks.
peytontolbert/action-transformer
A Transformer-based model for sequential decision-making tasks
peytontolbert/agent_kernel
verifier-driven coding agent with a strict TOLBERT context layer in front of the LLM policy.
peytontolbert/text-generation-lab
just a repo for my 100m model
peytontolbert/simple-moe
Simple MoE - Day 17 of 365 Days of Repos
peytontolbert/speech-generation-lab
Just a repo for my f5-tts distillation
peytontolbert/mobile-computer-use
Mobile Computer Use on App Store for iPhone chats of codex/cursor
peytontolbert/skills
Skills Catalog for Codex
peytontolbert/WebGL_Tools
Browser-based tools for building 3D environments, games, and demos quickly. Generated https://peytontolbert.com/showroom/ in a few hours.
peytontolbert/text-sims-llm
An intelligent autonomous character simulation that uses AI to create self-directed virtual beings. The character makes decisions, maintains needs, builds memories, and learns from experiences.
peytontolbert/tiny-self-improve-ai
a repository for minimal code self improving AI
peytontolbert/parameter-golf
Train the smallest LM you can that fits in 16MB. Best model wins!
peytontolbert/minimal-agent-kernel
minimal verifier-driven coding agent. Uses LLM-backed policy, two actions, deterministic verification, and selfplay-oriented task generation to solve bounded local software tasks and accumulate reusable trajectories.
peytontolbert/TOLBERT
TOLBERT is a hierarchical, ontology-structured variant of BERT that uses a Tree-of-Life representation to model multi-level, evolutionary or taxonomic relationships over large text corpora.